- From: <noah_mendelsohn@us.ibm.com>
- Date: Tue, 23 Sep 2003 17:29:21 -0400
- To: Tim Bray <tbray@textuality.com>
- Cc: Norman Walsh <Norman.Walsh@Sun.COM>, www-tag@w3.org
It seems to me that, as with the Metadata in URI discussion, it may be worthwhile to distinguish what you must not do, what you should not do, and what you should avoid if possible. Tim Bray writes: >> For example, in mailto: URIs, the bit >> before the @ is opaque, the bit after >> isn't at all. HTTP URIs are explicitly >> opaque after the host part. Accepting without debate that the above is in some sense true, I think it still pays to discourage people and software from depending on any aspect of the URI's particulars, including the scheme, except where necessary. I believe the folllowing is a quote from the metadata in URI document [1] (or maybe it's text that I proposed :-)...I'm on an airplane at the moment and can't get to the original, only some old email discussing it.) "People and software using URIs assigned outside of their own authority should make as few inferences as possible about a resource based on its identity and inferences arising from delegated authority. The more dependencies a piece of software has on particular constraints and inferences, the more fragile it becomes to change and the lower its generic utility." I believe that this is good advice, regardless of whether the http and/or mailto specs licenses inferences about the DNS names used. For me, the Metadata in URI did a good job of balancing these concerns. I learned a lot from reading it, and I suggest that its formulation be considered in clarifying the current discussion of opacity. Thank you! Noah [1] http://www.w3.org/2001/tag/doc/metaDataInURI-31 ------------------------------------------------------------------ Noah Mendelsohn Voice: 1-617-693-4036 IBM Corporation Fax: 1-617-693-8676 One Rogers Street Cambridge, MA 02142 ------------------------------------------------------------------
Received on Tuesday, 23 September 2003 17:37:06 UTC